Advanced methods for digital quadrature and offset modulation classification

Abstract

We address the problem of constructing classifiers of digital quadrature and offset modulated signals b y the likelihood approach and the Mth-law approach. In the first approach, we start from the likelihood functionals (LF) of such signals in additive white Gaussian noise (AWGN) and, based on these LFs, we derive easily implementable classifiers. In the second approach, Mth-law classifiers are examined for classifying such signals. It is shown that certain versions of the LF classifier are closely related to the Mth-law classifier in terms of both implementation and performance, at least in the low-SNR regime.

A qHLRT Modulation Classifier with Antenna Array and Two-Stage CFO Estimation in Fading Channels

“…Moreover, normally in practice the probability density function (pdf) of the unknown parameters is not available. Even further, averaging over these parameters to obtain the average likelihood function is quite difficult [2], [4]- [6], whereas ML estimation of them suffers from high computational complexity as well [7].…”